Unverified50% confidenceBenchmarkExact time
On 8 AMD Instinct MI355X accelerator cards, using the vLLM inference framework, the GLM-5.3 model achieved a single-user decoding speed of 469 tok/s.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/4/2026
First Seen
Valid until: 10/18/2026
Sources
Related Entities
Related Claims
UnverifiedThe AMD MI355X running GLM5.2 achieved an inference throughput of 2626 tokens/second on a single node.74% similarUnverifiedAMD and EmbeddedLLM jointly published a blog documenting the process of optimizing MiniMax M3 inference serving on Instinct MI355X hardware.71% similarUnverifiedThe AMD MI300X is AMD's flagship accelerator for AI training and inference, featuring Chiplet packaging, integrated 192 GB HBM3 unified memory, and a peak BF16 compute performance of approximately 1.3 PFLOPS per card.69% similarUnverifiedThis benchmark was conducted on four model families — Gemma, Qwen, Kimi, and MiniMax — using AMD Instinct MI300X and MI355X hardware.68% similarVerifiedThe MI355X belongs to AMD's Instinct series of data center accelerator cards and is evolved from the CDNA architecture.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/972278API
curl https://kongchang.com/api/v1/knowledge/claims/972278MCP
get_claim(id=972278)